Postgres case insensitive for UID

[fogelito]

Support Postgres case-insensitivity and accent-insensitive ONLY on $id attribute unique index

Summary by CodeRabbit

• Chores
  • Updated PostgreSQL database collation settings and index configuration to enhance consistency in text handling and database performance.
  • Improved test infrastructure for exception handling validation.

[coderabbitai]

Walkthrough

PostgreSQL adapter collation definitions updated from utf8_ci to utf8_ci_ai with adjusted locale settings. COLLATE clauses added to indexed columns (_uid, _document) across shared and non-shared table configurations. Test exception handling refactored from expectException to try-catch blocks for duplicate detection scenarios.

Changes

Cohort / File(s)	Summary
PostgreSQL Collation & Index Updates src/Database/Adapter/Postgres.php	Updated collation definition from utf8_ci (locale 'und-u-ks-primary') to utf8_ci_ai (locale 'und-u-ks-level1'). Added COLLATE utf8_ci_ai to unique indexes on "_uid" and "_document" fields in both shared and non-shared table modes. Simplified index attribute handling by removing conditional COLLATE logic in createIndex for UNIQUE indexes. Adjusted variable naming in deleteIndex from $name to $collection.
Test Exception Handling tests/e2e/Adapter/Scopes/DocumentTests.php	Replaced expectException declarations with try-catch blocks in four test methods (testUniqueIndexDuplicate, testUniqueIndexDuplicateUpdate, testExceptionDuplicate, testExceptionCaseInsensitiveDuplicate) to assert DuplicateException on document creation/update failures.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related PRs

• utopia-php/database#661: Modifies Postgres index creation logic and index/column handling for spatial enhancements, overlapping with collation and index attribute changes.
• utopia-php/database#613: Adjusts Postgres uniqueness handling for _uid/_tenant indexes with related index-creation logic modifications.

Suggested reviewers

• abnegate

Poem

🐰 Collations shift like carrot gardens fair,
With utf8_ci_ai blooming everywhere,
Indexes dressed in COLLATE's bright dress,
Try-catch blocks put tests to the test,
Postgres hops on with indexed finesse! 🥕✨

Pre-merge checks and finishing touches

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Docstring Coverage	✅ Passed	Docstring coverage is 88.89% which is sufficient. The required threshold is 80.00%.
Title Check	✅ Passed	The title clearly and concisely summarizes the primary change, indicating that Utopia’s Postgres adapter now treats UID comparisons as case-insensitive, which aligns with the pull request’s objective to apply a utf8_ci_ai collation to the UID index. It avoids extraneous detail while accurately reflecting the core enhancement.

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch postgres-case-insensitive

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 1

🧹 Nitpick comments (1)

tests/e2e/Adapter/Scopes/DocumentTests.php (1)

5331-5342: Add accent‑insensitive duplicate check for $id (Postgres‑only)

To fully cover the PR objective, also assert that 'café' vs 'cafe' (and/or 'CAFÉ') duplicates on $id under Postgres.

Example test to add (scoped to Postgres):

public function testExceptionAccentInsensitiveDuplicateUID(): void
{
    /** @var Database $database */
    $database = static::getDatabase();

    // Only relevant for the Postgres adapter
    if (!($database->getAdapter() instanceof \Utopia\Database\Adapter\Postgres)) {
        $this->expectNotToPerformAssertions();
        return;
    }

    $doc = new Document([
        '$id' => 'café',
        '$permissions' => [Permission::read(Role::any())],
    ]);

    // First insert
    $doc->removeAttribute('$sequence');
    $database->createDocument('documents', $doc);

    // Same ID without accent should conflict under accent-insensitive collation
    $doc->setAttribute('$id', 'cafe');
    $doc->removeAttribute('$sequence');

    try {
        $database->createDocument('documents', $doc);
        $this->fail('Failed to throw exception');
    } catch (\Throwable $e) {
        $this->assertInstanceOf(DuplicateException::class, $e);
    }
}

📜 Review details

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 1ce7ffc and 6b932c1.

📒 Files selected for processing (2)

• src/Database/Adapter/Postgres.php (5 hunks)
• tests/e2e/Adapter/Scopes/DocumentTests.php (4 hunks)

🧰 Additional context used

🧠 Learnings (1)

📚 Learning: 2025-10-03T02:04:17.803Z

Learnt from: abnegate
PR: utopia-php/database#721
File: tests/e2e/Adapter/Scopes/DocumentTests.php:6418-6439
Timestamp: 2025-10-03T02:04:17.803Z
Learning: In tests/e2e/Adapter/Scopes/DocumentTests::testSchemalessDocumentInvalidInteralAttributeValidation (PHP), when the adapter reports getSupportForAttributes() === false (schemaless), the test should not expect exceptions from createDocuments for “invalid” internal attributes; remove try/catch and ensure the test passes without exceptions, keeping at least one assertion.

Applied to files:

• tests/e2e/Adapter/Scopes/DocumentTests.php

🧬 Code graph analysis (2)

tests/e2e/Adapter/Scopes/DocumentTests.php (5)

src/Database/Adapter/Postgres.php (2)

• createDocument (950-1050)
• create (135-162)

src/Database/Adapter/MariaDB.php (2)

• createDocument (822-933)
• create (28-43)

src/Database/Adapter/SQLite.php (2)

• createDocument (522-630)
• create (113-116)

src/Database/Database.php (2)

• createDocument (4176-4270)
• create (1241-1260)

src/Database/Document.php (3)

• Document (12-470)
• getId (63-66)
• removeAttribute (287-293)

src/Database/Adapter/Postgres.php (3)

src/Database/Adapter/SQLite.php (1)

• getSQLTable (1080-1083)

src/Database/Adapter/SQL.php (1)

• getSQLTable (1704-1707)

src/Database/Adapter.php (1)

• filter (1196-1205)

⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)

• GitHub Check: Setup & Build Docker Image

🔇 Additional comments (7)

tests/e2e/Adapter/Scopes/DocumentTests.php (3)

4737-4765: LGTM: duplicate on unique index asserted correctly

Try/catch pattern and DuplicateException assertion look good and consistent with the suite.

---

4801-4807: LGTM: update collision handled

Update path correctly asserts DuplicateException via try/catch.

---

5309-5319: LGTM: preparing re-insert by clearing $sequence

Clearing '$sequence' before each create avoids manual sequence leakage; duplicate $id is asserted properly.

src/Database/Adapter/Postgres.php (4)

247-247: LGTM! Collation correctly applied to _uid indexes.

The COLLATE clause is properly added to the unique indexes on the _uid column for both shared and non-shared table modes, ensuring case-insensitive and accent-insensitive matching for the $id attribute as intended by the PR objectives.

Also applies to: 254-254

---

855-855: LGTM! Appropriate simplification of index attribute formatting.

The removal of per-attribute COLLATE logic from createIndex is correct, as the collation is now explicitly applied only to the built-in indexes on _uid and _document columns in createCollection. This makes the behavior more explicit and aligns with the PR objective to limit the collation changes to the $id attribute rather than all unique indexes.

---

903-903: LGTM! Improved variable naming for clarity.

Renaming the variable from $name to $collection improves code readability by using a more descriptive identifier.

---

282-282: Verify whether COLLATE on _document should be applied to shared mode permissions index.

The asymmetry is confirmed: in non-shared mode (line 282), the _document column in the unique permissions index gets COLLATE utf8_ci_ai, while in shared mode (lines 273–274), the _document column in the unique index has no COLLATE clause. No inline comments explain this difference.

Both modes use _document in the unique index structure, so this difference may be intentional or an oversight. Confirm whether shared mode should also apply COLLATE utf8_ci_ai on _document for consistency, or document the architectural reason for the asymmetry.